Governments have declared several large-scale, oceanic, marine protected areas (MPAs) in the Pacific Ocean in recent years ( Supplementary Table 1 ), while others are currently under consideration (Govan, 2020). Among the established MPAs, the Phoenix Islands Protected Area (PIPA), the Palau National Marine Sanctuary (PNMS) and the Pacific Remote Islands Monument (PRIM), occur in prime tropical tuna fishing grounds and are designated as no-take reserves in which all forms of extractive fishing are prohibited.

The proponents of large-scale, oceanic MPAs, particularly those in the tropical Pacific, often cite the benefits to tuna conservation, inter alia, as a major rationale for their establishment. For example, in respect of the PIPA it has been claimed that “the reserve is a major spawning ground for tuna, so its closure will have a major contribution to the conservation and rejuvenation of fish stocks and to global food security”, and “closing PIPA would be the single most effective act of marine conservation in history and a big step in preventing the world’s last major population of skipjack tuna from becoming as depleted as those of the Atlantic and Indian Oceans” (Oceana, 2016).

MPAs such as PIPA and PNMS are areas in which tropical tuna species, skipjack (Katsuwonus pelamis), yellowfin (Thunnus albacares) and bigeye (T. obesus) tuna, frequently spawn (Hernández et al., 2019), and hence it has been often assumed that MPAs will offer protection to spawning populations. Other potential benefits, such as spill-over of enhanced populations within the MPA into adjacent fished areas are also frequently cited (e.g., PICRC and COS, 2019; Medoff et al., 2022).

Despite claims that large, oceanic MPAs will benefit tuna stocks, their value, in common with spatial management approaches more broadly (Hilborn, 2018; Hilborn et al., 2021), to tuna conservation and to adjacent fisheries have not been quantitatively evaluated using realistic assumptions regarding tuna population dynamics and spatial connectivity. Tunas spawn widely throughout the tropical Pacific, their larvae are redistributed by ocean currents, and they develop habitat-driven movement capability as they grow beyond the first few months of life (Lehodey et al., 2008). The species biology, particularly the dispersal and movement characteristics, and the distribution of the population and fisheries in relation to the size and location of the MPAs will influence MPA efficacy as a fishery management and conservation tool (Gaines et al., 2010; Fulton et al., 2015; Ovando et al., 2021). Therefore, these characteristics need to be considered to understand the likely contribution of MPAs to tuna conservation and management.

The aim of this paper was to evaluate the conservation efficacy for two important and contrasting tuna species, skipjack and bigeye tuna, of an existing MPA, the PIPA, and a series of large hypothetical MPAs each constituting approximately 33% of the western and central Pacific Ocean (WCPO, Pacific Ocean west of 150°W, Figure 1 ). The evaluation was designed specifically to take account of the spatial population and fishery dynamics of these species as well as the size and location of the MPAs.

We chose the PIPA as the main case study for this paper because it is located within the core habitat of tropical tuna in the WCPO ( Figure 1 ) and, prior to its declaration, had annual catches of skipjack tuna of up to 110,000 t (6% of the total WCPO catch) and of bigeye tuna of up to 5,000 t (3% of the WCPO catch; Figure 2 ). However, the PIPA comprises only 0.6% of the area of the WCPO between the latitudes 40°N and 40°S (the approximate limits of skipjack and bigeye tuna distribution). As a result, the PIPA might not be expected, a priori, to be highly impactful on these tuna stocks over their entire range. Therefore, we also investigated the efficacy of three much larger, hypothetical closures, each representing approximately 33% of the WCPO ( Figure 1 ). Closures of this magnitude are currently being advocated globally (e.g., Roberts et al., 2019) to protect biodiversity and the abundance of fished populations.

Skipjack tuna constitutes two-thirds of the total WCPO tuna catch (3 million t in 2019; Hare et al., 2021), about 40% of the global tuna catch (WCPFC, 2020), is the key target species of the large tropical purse seine and a smaller pole-and-line fishery and is critical for the economic and social well-being of Pacific Islanders (Gillett, 2016). Skipjack is a small (maximum size ~90 cm fork length), fast-growing and relatively short-lived (~4 yr.) tuna species thought to be highly resilient to fishing and capable of sustaining current annual catches in the WCPO approaching 2 million t (Vincent et al., 2019). By contrast, bigeye tuna is larger (maximum size ~180 cm fork length in the WCPO), longer lived (>10 yr.), supports considerably smaller annual catches in the WCPO of <200,000 t and is more heavily exploited than skipjack tuna (Ducharme-Barth et al., 2020; Hare et al., 2021). Bigeye tuna is the main target species of the tropical longline fishery, which targets larger fish sizes for the sashimi market. Also, small juvenile bigeye tuna are caught in the tropical purse seine and other surface fisheries targeting skipjack tuna. The importance and contrasting population dynamics and exploitation of these two species in the WCPO motivated their selection as case studies to evaluate the contribution of spatial closures such as PIPA, and the much larger hypothetical 33% closures, to the conservation of tuna and similar species in the Pacific region.

This work represents the first quantitative evaluation of the conservation benefits for two major, commercially important tuna species with contrasting demographic and exploitation characteristics, skipjack and bigeye tuna, of an actual spatial closure, the PIPA, in the Pacific region. The analysis shows that the conservation benefits for skipjack tuna within the MPA, near-MPA and stock-wide scales are negligible. Benefits for bigeye tuna are slightly greater, but not of a magnitude that would make a meaningful contribution to conservation. These results span a 22-year period comprising considerable environmental variation, including strong El Niño and La Niña events. El Niño conditions have been shown to expand the western Pacific warm pool (sea surface temperature >28.5°C) and skipjack tuna distribution from the tropical western Pacific into the tropical central and eastern Pacific (Lehodey et al., 1997). The 1998 and 2015-2016 skipjack tuna SB peaks in the central-Pacific-located PIPA ( Figure 3B ) and P-5deg areas ( Figure 3C ) are likely related to the strong El Niño events that occurred around those times (Hanich et al., 2018).

The results pertaining to the PIPA, which are broadly consistent with an empirical study of the impacts of a similar sized MPA (the British Indian Ocean Territory) on bigeye and yellowfin tuna in the Indian Ocean (Curnick et al., 2020), are not surprising. The PIPA is a small area relative to the distribution of the highly mobile skipjack (Sibert and Hampton, 2003) and bigeye tuna (Schaefer et al., 2015) stocks. Furthermore, although both species spawn in the PIPA (Hernández et al., 2019), they also spawn broadly in the tropical WCPO ( Figures 4A , 5A ) and are subject to current velocities that result in high levels of passive drift (Scutt Phillips et al., 2019). The PIPA on average is estimated by SEAPODYM to supply only 1.3% and 0.9% of the total WCPO reproductive output for skipjack and bigeye tuna, respectively ( Supplementary Table 4 ). It is therefore clear that some expectations for the impact of PIPA on tuna conservation (e.g., Oceana, 2016) were overly optimistic. The Kiribati Government has recently decided to re-open the PIPA to controlled commercial fishing. The results presented here show that there would be no consequential negative conservation impacts on skipjack or bigeye tuna resulting from the implementation of this decision.

Our results also show that even for much larger hypothetical closures comprising approximately 33% of the WCPO area, conservation impacts on skipjack tuna would likely be modest, with average changes in SB of up to +5.8% across the WCPO. Larger increases were estimated in the zones assumed to be closed, e.g., +23.5% for the western zone and +16.1% for the central zone; however, these need to be counter-balanced by reductions in SB of up to -13.8% in the zones remaining open, resulting from the additional concentration of fishing activity transferred from the closed zones. The results for the more heavily exploited bigeye tuna show higher conservation benefits, with average changes in WCPO SB of +4.8% to +12.0%.

We conclude that large-scale oceanic MPAs such as the PIPA or even the much larger spatial closures, generate very modest increases in stock-wide biomass of skipjack tuna, which is currently subject to moderate exploitation pressure, or even the more heavily exploited bigeye tuna at recent historical levels of exploitation. The conservation benefits that were estimated, even for the large spatial closures, were not of a magnitude to suggest that they could be effective frontline management tools for tuna in the WCPO. Note that our estimates pertain explicitly to the protective effects of MPAs in providing fishing refugia within their boundaries rather than any secondary effects that MPAs might have on modifying total, stock-wide levels of fishing. Various fishery controls, such as catch and effort limits and seasonal closures of purse seine sets on fish aggregation devices, have been implemented in the WCPO since 2010 (WCPFC, 2022) and are reflected in both the Fref and counterfactual simulations.

The reasons for the lack of effectiveness of spatial closures are rooted in the population dynamics of species such as skipjack and bigeye tuna, which have wide distributions and are highly mobile (Hilborn et al., 2021). This results in fishing impacts, even from a spatially restricted source, being dispersed to some extent throughout the range of the stock. The corollary is that MPAs can only provide protection from fishing when the fish are present in the MPA, and these protective gains are dissipated when fish move beyond the limits of the MPA. Additionally, MPAs do not control the level of stock-wide fishing activity, but typically move it into that part of the fishery that remains open (Sibert et al., 2012; Hilborn et al., 2021). As shown in the R33 scenarios here, large spatial closures that crowd fishing effort into the remaining two-thirds of the fishing grounds, could reduce local abundance in those areas and result in decreased economic performance of fishing vessels – lower per-vessel catches and catch-per-unit-effort. Proposals to close the high seas to fishing (White and Costello, 2014; Sumaila et al., 2015) maybe also produce negative effects of this nature. Unless carefully designed, large spatial closures may also impact negatively on onshore processing and service facilities in developing economies, whose competitive advantage substantially relies on proximity to fishing grounds for vessels supplying and/or using those facilities. The R33 closure scenarios that we have examined here are naïve in the sense that they were used only for demonstration purposes and ignored the reality that large closures of this sort would be difficult if not impossible to implement in a politically acceptable and equitable way. Similarly, while “marine corridors” or spatiotemporally dynamic MPAs have been proposed to increase effectiveness for highly mobile species (O’Leary et al., 2018; Boerder et al., 2019; Gilman et al., 2019), their near real-time implementation at scale in oligotrophic, tropical zones are likely to be even more problematic. In the case of tuna fisheries, more direct control of catch and effort through a multilateral regional fisheries management organization such as WCPFC is likely to be a more effective and equitable management tool and result in less social upheaval, particularly for nearshore tuna fisheries that contribute to food security in places like the Pacific Islands and Southeast Asia.

These conclusions are based on high-resolution spatial models of skipjack and bigeye tuna population dynamics and fisheries. The models assimilate current understanding of stock demographics and available fishery and conventional tagging data through the parameter estimation process (Senina et al., 2020; Senina et al., 2021). The models therefore provide the best current representation of age-specific movement, reproduction, mortality, exploitation, and the environmentally driven seasonal, inter-annual and spatial distribution of abundance, which is critical for the evaluation of spatial management schemes.

However, no model is perfect, and these are no exception. Some caveats of our modelling approach to bear in mind are: (i) we are modelling the dynamics of single pelagic species, assumed to be a single, self-replenishing stock and with no consideration of species interactions through competition for prey or inter-species predation; (ii) we do not model the impact on prey species of predation by the modelled pelagic species; and (iii) the model results are conditioned on the oceanographic forcing reanalyses used in the simulations, and while these are the best currently available, it is possible that they contain biases in some areas. We have not attempted to estimate sensitivity of our results to these and other possible sources of uncertainty. However, the approach taken for the evaluation of comparing retrospective reference and counterfactual simulations (Fulton et al., 2015) would tend to reduce the risk of such uncertainties impacting our conclusions as any errors or bias would tend to affect both simulations in a similar way. The retrospective approach, compared to the alternative of attempting to project into the future, has the added advantage of working with known historical fishery distributions and oceanic environmental properties that are required to drive the simulations. The period covered by the study (1998-2019) encompasses the full range of typically observed environmental variability in the Pacific, including several strong El Niño and La Niña events, thus capturing any interactions between spatial management effectiveness and environmental variability.

Our findings have contrasted two species of tropical tuna with differing population dynamics and degrees of fishing-induced biomass depletion, and should also be informative regarding the likely effects of MPAs on other widely distributed, mobile pelagic species such as yellowing tuna, whose population and exploitation characteristics lie somewhere between skipjack and bigeye tuna. However, our examples do not include species whose biomass has been depleted to levels below biological sustainability limits or stocks that systematically migrate to discrete spawning grounds. How MPAs might contribute to the conservation of pelagic stocks with these characteristics remains untested. No tropical tuna stocks in the WCPO have been depleted to levels below their biological sustainability levels (Hare et al., 2021). Similarly, while stock structure for tropical tuna stocks in the Pacific remains uncertain (Moore et al., 2020), the most recent information from genetic studies has not detected strong evidence of fidelity to discrete spawning grounds (Grewe et al., 2015; Anderson et al., 2020; Natasha et al., 2022).

We recognise that large oceanic MPAs such as the PIPA may have conservation value in areas not analysed in this paper, for example in the protection of coral reef ecosystems and their resident species that might otherwise be heavily fished. However, where the objectives of, and justification for, MPAs include enhancing the conservation of widely distributed pelagic species, this should be carefully evaluated during the planning stage, for example using methods such as those of Ovando et al. (2021) and that presented in this paper. This would allow expectations to be realistic and evaluated against the likely economic and social costs and benefits that would occur.

The data analyzed in this study is subject to the following licenses/restrictions: Fisheries data used in this study are held by the Pacific Community on behalf of its members. Such data are available to third parties with the approval of the owners of the data. Reasonable requests for access to data will be facilitated by the Pacific Community. Environmental data used in the SEAPODYM models are publicly available and can be provided on request to the authors. Requests to access these datasets should be directed to Dr John Hampton, johnh@spc.int.

JH, PL and IS designed and contributed equally to the study, IS developed the SEAPODYM models and the associated high-resolution models used in the simulations, PL conducted the simulations, JH and PL wrote the paper, with contributions and review by IS, SN, JSP and KT. KT provided information on the PIPA and relevant fisheries management context.